Open Access N→ogNzs3gN→ooo3 Figure 3 Double-hierarchical clustering analysis. The faecal samples from 20 patients at three time points for 36 dominant the samples at the family level. Sample ID numbers 1-20, 21-40 and 41-60 on the left side represent those obtained at Sz o families are represented on a double-hierarchical clustering heat map. The clusters on the right side indicate similarity amor S2 and S3, respectively. IDs 1, 21 and 41, and likewise 2, 22 and 42 and so on correspond to the same patient. The clusters at the bottom indicate the families showing similarity in the frequency of identification among samples. the Z-score column represents the relative ratio of abt an opposite pattern of colonisation was found between In the present study, the gut microbiota after the eradi Bifidobacteriaceae and Ruminococcaceae whose major genera cation therapy showed a phenotype not of obesity but of 9 That is, a sample with a high density (columns with ratio and an abundance of E prausnitzii. Furthermore, the 9 red colour) of Bifidobacteriaceae contained a low density plasma level of the active form of ghrelin was significantly Columns with blue colour) of Ruminococcaceae, and vice lower after eradication therapy than before the therapy. A common niche in the gut ecosystem. Thus, the adminis- between such changes in the B: F ratio and the ghrelin oo tration of both genera together as probiotics may result in level. The majority of circulating ghrelin is produced in the x competition for nutrition and colonisation sites, despite gastric mucosa and stimulates the appetite, thus increasing their being considered beneficial bacteria for improving the fat mass deposition and body weight. 0 Therefore, in A energy balance 20 21 Indeed, Faecalibacterium is known energy homeostasis, the increase in the B F ratio in the gut butyrate in the human gut microbiota &g-s that produce microbiota was comparable with the decrease in the plasma to be one of the major bacterial group ghrelin level. It has been reported that the plasma ghrelin a Yanagi H, et al. BM Open Gastro 2017: 4: 0000182. doi: 10. 1136/bmigast-2017-0001826 Yanagi H, et al. BMJ Open Gastro 2017;4:e000182. doi:10.1136/bmjgast-2017-000182 Open Access an opposite pattern of colonisation was found between Bifidobacteriaceae and Ruminococcaceae whose major genera were Bifidobacterium and Faecalibacterium, respectively. That is, a sample with a high density (=columns with red colour) of Bifidobacteriaceae contained a low density (=columns with blue colour) of Ruminococcaceae, and vice versa. This suggests that these two genera may share a common niche in the gut ecosystem. Thus, the adminis￾tration of both genera together as probiotics may result in competition for nutrition and colonisation sites, despite their being considered beneficial bacteria for improving energy balance.20 21 Indeed, Faecalibacterium is known to be one of the major bacterial groups that produce butyrate in the human gut microbiota.22 In the present study, the gut microbiota after the eradi￾cation therapy showed a phenotype not of ‘obesity’ but of ‘lean’, which was represented by an increase in both the B:F ratio9 and an abundance of F. prausnitzii. 22 Furthermore, the plasma level of the active form of ghrelin was significantly lower after eradication therapy than before the therapy. A statistical analysis demonstrated a significant correlation between such changes in the B:F ratio and the ghrelin level. The majority of circulating ghrelin is produced in the gastric mucosa and stimulates the appetite, thus increasing the fat mass deposition and body weight.3 10 Therefore, in energy homeostasis, the increase in the B:F ratio in the gut microbiota was comparable with the decrease in the plasma ghrelin level. It has been reported that the plasma ghrelin Figure 3 Double-hierarchical clustering analysis. The faecal samples from 20 patients at three time points for 36 dominant families are represented on a double-hierarchical clustering heat map. The clusters on the right side indicate similarity among the samples at the family level. Sample ID numbers 1–20, 21–40 and 41–60 on the left side represent those obtained at S1, S2 and S3, respectively. IDs 1, 21 and 41, and likewise 2, 22 and 42 and so on correspond to the same patient. The clusters at the bottom indicate the families showing similarity in the frequency of identification among samples. The Z-score column represents the relative ratio of abundance. copyright. on 4 July 2018 by guest. Protected by http://bmjopengastro.bmj.com/ BMJ Open Gastroenterol: first published as 10.1136/bmjgast-2017-000182 on 26 November 2017. Downloaded from